Why a reed switch needs no power of its own
The whole mechanism is two blades of nickel iron sealed in glass, overlapping but not touching, with the gap set during manufacture. A magnetic field along the axis of the tube makes one blade a north pole and the other a south, they attract, and the contacts close. Remove the field and the springiness of the blades opens them again. There is no coil, no oscillator and no electronics, so there is nothing to supply.
That is the practical difference from an inductive or capacitive proximity sensor, which needs a supply voltage to run its oscillator and gives you a transistor output. A reed switch is a genuine dry contact: you can put it in series with a bell, a relay coil or a PLC input without caring which side is positive, and it will still work if the rest of the panel is unpowered.
The glass envelope is filled with nitrogen or is evacuated, which is why the contacts stay clean for tens of millions of operations in a dry, non-corroding atmosphere. It is also the vulnerability: glass breaks, so the bare switches are usually potted into a plastic body or a threaded barrel before they meet the real world.
Form A, Form B and Form C
Form A is the default and the one most listings mean by normally open: no magnet, no circuit. Bring the magnet near and it closes. This is what goes on a door or window, so that opening the door takes the magnet away and breaks the loop.
Form B is normally closed, and it is worth understanding how it is achieved, because a reed switch cannot be normally closed on its own. A small bias magnet is fitted alongside the blades to hold them together, and the actuating magnet is arranged to cancel that bias so the contacts open. That makes Form B slightly more sensitive to how the magnet is presented, and it is why a normally closed unit costs more than a normally open one.
Form C is a changeover with a common, a normally open and a normally closed contact in the same envelope. Listings that offer normally open and normally closed as a choice are usually selling two different parts, whereas a genuine Form C gives you both from one device and three wires.
| Form | Behaviour with no magnet | Typical use |
|---|---|---|
| Form A | Open | Door and window contacts, position sensing, flow paddles |
| Form B | Closed, held by a bias magnet | Fail-safe loops where a cut wire must raise the alarm |
| Form C | Changeover, common plus NO and NC | Where one sensor has to switch something on and something off |
Contact rating is the number that kills them
Reed switches are rated three ways and all three matter: a maximum switching voltage, a maximum switching current, and a maximum switching power in watts or volt amperes. The power figure is the binding one. A switch quoted at 200 V and 0.5 A with a 10 W maximum cannot do 200 V and 0.5 A at the same time, because that is 100 W. Work out the actual load and check it against the power figure, not just the voltage.
The failure that catches people is inrush rather than steady current. A filament lamp draws many times its running current for the first few milliseconds, and a capacitive load, including the input of many LED drivers and switched mode supplies, does the same. That spike arrives on contacts that are still bouncing, and it welds them. The switch then reads permanently closed and looks like a fault in whatever it was controlling.
The fix is not a bigger reed switch, it is to stop switching the load directly. Put the reed switch on a relay coil, a PLC input or a transistor gate and let something built for the job carry the current. Where the load must be switched directly, a series resistor or a small inductor to slow the inrush, and a snubber across an inductive load, are the usual remedies. A reed relay is the packaged version of the same idea: a reed switch inside a small coil, giving isolation and a fast, bounce-free contact for signal switching.
Sensing distance, hysteresis and pneumatic cylinders
Sensitivity is quoted in ampere turns, and in practice what you care about is the pull-in distance, where the contacts close as the magnet approaches, and the drop-out distance, where they open again as it leaves. Drop-out is always further away than pull-in, and that gap is useful hysteresis: it stops the contacts chattering when a door rattles or a piston creeps.
Orientation matters as much as distance. The field has to run along the axis of the glass tube, so a magnet presented broadside gives a much shorter working range than the same magnet presented end on, and two magnets pushed close together can cancel rather than reinforce. If a sensor works on the bench and not on the machine, the usual cause is the approach angle rather than a weak magnet.
On pneumatic cylinders the reed switch is the standard piston sensor, sold in barrel and T-slot forms with designations such as D-A93 or CS1-G. It works because the piston carries a ring magnet and the cylinder tube is aluminium, which the field passes straight through. The two wire versions with an LED are polarity sensitive and include a series resistor for their indicator, so they need to be wired the right way round even though a bare reed switch does not.
